Linkages between Forestry Best Management Practices and erosion in the southeastern US

Numerous studies have concluded that forestry Best Management Practices (BMPs) are effective at mitigating erosion and sedimentation caused by forest operations; however, the complex relationship between forestry BMPs and erosion is largely unexamined. In this study, BMP implementation rates, which are percentages ranging from 0 to 100% that indicates how well an operator instituted recommended practices in the field, and predicted erosion rates, obtained by using USLE-Forest, were calculated for 108 recent harvests in twelve states and three physiographic regions in the southeastern U.S. BMP implementation rates were subdivided into three levels of application: BMP+ (>90% implementation), BMP-standard (80-90% implementation), and BMP- (<80% implementation). Skid trails (86.5 Mg ha(-1)yr(-1)) and haul roads (90.3 Mg ha(-1) yr(-1)) eroded at relatively high rates at the BMP- level across the southeast. This emphasizes the importance of adequate BMP measures such as utilizing water diversion structures and cover management at these features to better protect water quality. The overall weighted average erosion estimates for all regions at the BMP-standard (10.4 Mg ha(-1) yr(-1)) and BMPT (6.6 Mg ha(-1) yr(-1)) levels were <11 Mg ha(-1) yr(-1), indicating that water quality and site productivity are largely protected when adequate BMPs are implemented, and Streamside Management Zones (SMZs) are utilized along streams. Approximately 94% of the sites sampled were classified as either BMP-standard or BMPT, demonstrating that BMPs are being implemented consistently throughout the southeast. Spearman p correlation analyses were performed for all variables. Forestry BMP implementation and erosion estimates had significant negative correlations, especially for skid trails (Spearman rho = -0.59, p-value < 0.0001) and haul roads (Spearman rho = -0.39, p-value = < 0.0001), as well as for all regions across the southeast. These variables, however, were poorly correlated for stream crossings, indicating that current audit questions in the southeast may not fully address erosion. Additionally, BMP implementation and erosion estimates exhibited a significant negative correlation (R-2 = 0.28, p-value < 0.0001) based on a quadratic regression line for all features, reinforcing that as BMP implementation increases, predicted erosion generally decreases.